Abstract: One or more portions of a text corpus indicative of operational risk from a set of risk assessment documents is identified. Contextual features from the one or more portions of the text corpus are determined by applying a natural language processing (NLP) algorithm on the one or more portions. Risk identifiers are extracted based on the determined contextual features. A risk assessment software is generated based on the extracted risk identifiers and an operational risk category.

Abstract: An approach is provided that receives a message and applies a deep analytic analysis to the message. The deep analytic analysis results in a set of enriched message embedding (EME) data that is passed to a trained neural network. Based on a set of scores received from the trained neural network, a conversation is identified from a number of available conversations to which the received message belongs. The received first message is then associated with the identified conversation.

Abstract: An approach is provided that receives a message and applies a deep analytic analysis to the message. The deep analytic analysis results in a set of enriched message embedding (EME) data that is passed to a trained neural network. Based on a set of scores received from the trained neural network, a conversation is identified from a number of available conversations to which the received message belongs. The received first message is then associated with the identified conversation.

Abstract: Clustering a set of natural language queries NLQs based on a set of significant events retrieved from a corpus stored in a computer system is described. A set of NLQs is used by a search engine for searching a selected corpus to retrieve respective sets of significant events. The set of NLQs is clustered into a plurality of NLQ clusters according to a number of common significant events being returned by the search engine for respective members of an NLQ cluster.

Abstract: In response to running at least one testing phrase on a previously trained text classifier and identifying a separate predicted classification label based on a score calculated for each respective at least one testing phrase, a text classifier decomposes extracted features summed in the score into word-level scores for each word in the at least one testing phrase. The text classifier assigns a separate heatmap value to each of the word-level scores, each respective separate heatmap value reflecting a weight of each word-level score. The text classifier outputs the separate predicted classification label and each separate heatmap value reflecting the weight of each word-level score for defining a heatmap identifying the contribution of each word in the at least one testing phrase to the separate predicted classification label for facilitating client evaluation of text classification anomalies.

Abstract: A knowledgebase of an expert system is populated with rules inferred from a set of business processes that govern the manner in which the business interacts with users. Each business process contains an input, an output, an action, and a set of dependency relationships that relate pairs of the input, the output, and the action. Each process's input, output, action, and dependency relationships are translated, respectively, into a subject, an object, a predicate, and a set of dependency relationships among the subject, object, and predicate, of a natural-language rule. Each rule is stored in the expert system's knowledgebase as a directed graph, and nodes representing each stored subject, object, and predicate are assigned domain classifications as a function of characteristics of the business rule. These domain classifications are represented within the knowledgebase as a set of domain classifications determined as a further function of characteristics of the business rule.

Abstract: Embodiments provide a computer implemented method for generating a domain-specific thesaurus on a cognitive system, comprising: receiving data of the domain-specific corpus and a plurality of terms of interest from a user; splitting the data of the domain-specific corpus into a plurality of sentences using natural language processing techniques; for each term in the plurality of terms of interest, retrieving a plurality of candidate sentences containing a corresponding term, from the plurality of sentences; for each candidate sentence, providing a list of synonyms of the corresponding term, wherein the synonyms are contextual alternatives in the corresponding candidate sentence; for each term in the plurality of terms of interest, tracking a frequency of each synonym, and forming a frequency map including all the synonyms of a corresponding term and the frequency of each synonym; and generating a domain-specific thesaurus based on a combination of all the synonyms in the frequency map.

Abstract: A method for providing discovery and realization of business measurement concepts may include providing at least one interface configured to receive an input from an operator associated with an organization, determining a selected set of glossary terms from a repository including a plurality of glossaries based at least in part on the input where the glossaries relate to different performance indicator components that are combinable to define a measurable performance indicator, and generating at least one performance indicator of the organization based on the selected set of glossary terms.

Abstract: A search space can be reduced using a pruned ontology that comprises entities from an initial ontology. A document corpus having a plurality of documents is received for processing. An ontology pruner determines a set of entities in the plurality of documents. For each entity in the set of entities, the ontology pruner determines a link for the entity in the initial ontology. The ontology pruner determines a score for the entity based on information obtained through the link for the entity in the initial ontology. The ontology pruner omits the entity from the pruned ontology if the score is below a predetermined or configurable threshold value.

Abstract: According to embodiments of the present invention, medical treatment outcomes are simulated. A system receives a request to determine a medical treatment pertaining to a medical condition of a patient. The request is applied to one or more models, via a processor, to simulate outcomes for a plurality of different medical treatments for the medical condition, wherein the one or more models account for impacts of the plurality of different medical treatments on use of medical treatments awaiting future approval. The medical treatment is determined from the simulated outcomes, via a processor, with a desired level of impact on the medical treatments awaiting future approval based on characteristics of the patient. Methods and computer readable media are also provided herein for simulating medical treatment outcomes.

Abstract: A test controller submits testing phrases to a text classifier and receives, from the text classifier, classification labels each comprising one or more respective heatmap values each associated with a separate word. The test controller aligns each of the classification labels corresponding with a respective testing phrase. The test controller identifies one or more anomalies of a selection of one or more classification labels that are different from an expected classification label for the respective testing phrase. The test controller outputs a graphical representation in a user interface of the selection of one or more classification labels and one or more respective testing phrases with visual indicators based on one or more respective heatmap values.

Abstract: A test controller submits testing phrases to a text classifier and receives, from the text classifier, classification labels each comprising one or more respective heatmap values each associated with a separate word. The test controller aligns each of the classification labels corresponding with a respective testing phrase. The test controller identifies one or more anomalies of a selection of one or more classification labels that are different from an expected classification label for the respective testing phrase. The test controller outputs a graphical representation in a user interface of the selection of one or more classification labels and one or more respective testing phrases with visual indicators based on one or more respective heatmap values.

Abstract: A technology that facilitates transparency of claim processing for a third-party payer is disclosed. Exemplary implementations may: receive a claim submission from a service provider and/or one or more agents thereof to seek settlement for goods and/or services provided to a beneficiary of a third-party payer; receive beneficiary data from the third-party payer; combine the claim-submission data and the beneficiary data to form claim data; store the claim data in an immutable and appendable format; obtain a determination of a state of the claim that is based at least in part on the claim data; update the claim data by appending the determination of the state of the claim; and provide access to the stored claim data to the third-party payer and the service provider and/or the one or more agents thereof.

Abstract: A knowledge graph is built based on a corpus stored in the computer system. The corpus includes a set of searchable events and each event includes a respective set of entities. A set of entities is identified in a first set of significant events returned by natural language query (NLQ). The knowledge graph determines which ones of the set of entities are related to the entities in the NLQ to produce a filtered set of entities. The filtered set of entities is used to identify a second set of significant events in the selected corpus. Members of the first and second set of significant events are presented to a user as a search result.

Abstract: Clustering a set of natural language queries NLQs based on a set of significant events retrieved from a corpus stored in a computer system is described. A set of NLQs is used by a search engine for searching a selected corpus to retrieve respective sets of significant events. The set of NLQs is clustered into a plurality of NLQ clusters according to a number of common significant events being returned by the search engine for respective members of an NLQ cluster.

Abstract: Clustering a set of natural language queries NLQs based on a set of significant events retrieved from a corpus stored in a computer system is described. A set of NLQs is used by a search engine for searching a selected corpus to retrieve respective sets of significant events. The set of NLQs is clustered into a plurality of NLQ clusters according to a number of common significant events being returned by the search engine for respective members of an NLQ cluster.

Abstract: Electronic natural language processing in a natural language processing (NLP) system, such as a Question-Answering (QA) system. A receives electronic text input, in question form, and determines a readability level indicator in the question. The readability level indicator includes at least a grammatical error, a slang term, and a misspelling type. The computer determines a readability level for the electronic text input based on the readability level indicator, and retrieves candidate answers based on the readability level.

Abstract: An engagement classifier for a group chatbot is trained by leveraging the implicit dataset generated by humans engaging in both direct messages as well as group conversations. Human-to-human direct messages are used as an approximate representation of the domain knowledge and expertise of each user. The decision to engage in a group conversation is assumed to be based on that domain knowledge. The knowledge representations and instances of engagements in group conversations yields an effective set of features and labels which can be used to model the engagement decision. The same transfer learning technique is used to generate a knowledge representation for the group chatbot. Given this representation of the domain knowledge of the chatbot, the classifier can predict whether it should engage in any particular group conversation.

Abstract: A method, system and computer-usable medium are disclosed for preserving temporal relevance of content within a corpus. A corpus is processed to generate temporally-relevant metadata and mined content, which in turn are processed to generate first temporal relevancy metrics. The cache history of a web browser is likewise processed to generate second temporal relevancy metrics, which in turn is processed with the first temporal relevancy metrics to generate first temporal relevancy scores. New documents are ingested into the corpus and existing documents are revised. Temporally-relevant metadata and mined content associated with the updated corpus are then processed to generate third temporal relevancy metrics. The second and third temporal relevancy metrics are then processed to generate second temporal relevancy scores, which is then used to provide a temporally-relevant response to a query.

Abstract: An approach is provided to receive, at a question answering (QA) system, a question and identify a politeness corresponding to a number of terms corresponding to the question that are included in a corpus of the QA system. The approach identifies the politeness of one or more terms included in each of a set of candidate answers responsive to the question. Finally, the approach scores each of the candidate answers, with the scoring being based, in part, on the politeness identified for each of the terms.